Human Pose Estimation on AIC (val)
34.8APViTPose++-H
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| ViTPose++-HBackbone=ViT-H, Resolution=256x1922022.12 | 34.8 | 80.2 | 24.5 | 39.1 | 82.2 | |
| ViTPose-HBackbone=ViT-H, Resolution=256x1922022.12 | 34.6 | 80.2 | 24.3 | 39 | 82.1 | |
| HRFormerBackbone=HRFomer-B, Resolution=256x1922022.12 | 34.4 | 78.3 | 24.8 | 38.7 | 80.9 | |
| ViTPose++-LBackbone=ViT-L, Resolution=256x1922022.12 | 34.3 | 79.1 | 24.1 | 38.9 | 81.8 | |
| ViTPose-LBackbone=ViT-L, Resolution=256x1922022.12 | 34.1 | 79.1 | 23.9 | 38.7 | 81.6 | |
| HRNetBackbone=HRNet-w48, Resolution=256x1922022.12 | 33.5 | 78 | 23.6 | 37.9 | 80 | |
| UniHCPGFLOPS=18.9, Backbone=ViT-Base, Input Resolution=256x1922025.05 | 32.6 | — | — | — | — | |
| HRNetBackbone=HRNet-w32, Resolution=256x1922022.12 | 32.3 | 76.2 | 21.9 | 36.6 | 78.9 | |
| PoseBHGFLOPS=18.5, Backbone=ViT-Base, Input Resolution=256x1922025.05 | 32.1 | — | — | — | — | |
| ViTPose++-BBackbone=ViT-B, Resolution=256x1922022.12 | 31.8 | 76.7 | 20.6 | 36.3 | 79 | |
| HRFormerBackbone=HRFomer-S, Resolution=256x1922022.12 | 31.6 | 75.9 | 20.9 | 35.8 | 78 | |
| ViTPose++GFLOPS=18.5, Backbone=ViT-Base, Input Resolution=256x1922025.05 | 31.6 | — | — | — | — | |
| ViTPose-BBackbone=ViT-B, Resolution=256x1922022.12 | 30.9 | 75.8 | 19.4 | 35.3 | 78.3 | |
| SimpleBaselineBackbone=ResNet-152, Resolution=256x1922022.12 | 29.9 | 73.8 | 18.3 | 34.3 | 76.9 | |
| ViTPose++-SBackbone=ViT-S, Resolution=256x1922022.12 | 29.7 | 74.6 | 17.6 | 34.3 | 77.1 | |
| SimpleBaselineBackbone=ResNet-101, Resolution=256x1922022.12 | 29.4 | 73.6 | 17.4 | 33.7 | 76.3 | |
| ViTPose-SBackbone=ViT-S, Resolution=256x1922022.12 | 28.2 | 72.4 | 15.8 | 32.4 | 75.1 | |
| SimpleBaselineBackbone=ResNet-50, Resolution=256x1922022.12 | 28 | 71.6 | 15.8 | 32.1 | 74.1 |